From Signal Quality to Representation Stability in Textile Bioelectronic Interfaces.

Textile bioelectronic interfaces are emerging as wearable platforms for continuous physiological sensing, rehabilitation, and human-machine interaction. By integrating functional materials into textiles, electrodes and sensors can become part of the garment itself, forming a deformable interface with the body. Using surface electromyography (sEMG) acquired through textile-integrated electrodes as an illustrative case, this perspective argues that measured sEMG signals are shaped not only by muscle electrical activity but also by the deformable body-textile-electrode interface, whose state may change during use. Current evaluation of textile bioelectronic interfaces is often output centered, focusing on signal quality and downstream performance, which are necessary but insufficient for assessing whether measured signal patterns remain stable and reliable outside controlled conditions. This perspective proposes representation stability as a complementary evaluation criterion within an interface-centered framework. Representation stability asks whether measured signal patterns remain comparable as the physical sensing interface changes during real use. Within this framework, interface conditions become explicit test variables, with representation stability assessed using pre-specified criteria. Introducing representation stability extends evaluation from an output-centered toward an interface-centered paradigm. This could provide a more systematic basis for understanding, comparing, and ultimately designing textile bioelectronic interfaces that remain reliable under real-use variability.